Establishing moderators and biosignatures of antidepressant response in clinical care (EMBARC): Rationale and design

Association of anger attacks with suicidal ideation in adults with major depressive disorder: Findings from the EMBARC study

BACKGROUND

This report evaluates whether anger attacks (sudden uncharacteristic bouts of anger that are associated with autonomic arousal and/or aggression) in patients with major depressive disorder (MDD) are associated with elevated suicidal ideation (SI; active suicidal thoughts and plans).

METHODS

Participants of Establishing Moderators and Biosignatures of Antidepressant Response in Clinical Care (EMBARC) study who completed Massachusetts General Hospital Anger Attack Questionnaire (AAQ) at baseline were included (n = 293). Levels of SI (suicidal thoughts factor of Concise Health Risk Tracking) were compared at baseline with generalized linear models, and during Stage 1 (baseline-to-week-8) and Stage 2 (week-8-to-week-16) with repeated-measures mixed model analyses. Covariates included age, sex, race, ethnicity, site, and treatment arm.

RESULTS

At baseline, participants with (n = 109) versus without anger attacks (n = 184) had higher levels of SI (Cohen's d effect size [d] = 1.20). Those with ≥9 anger attacks in the past month had significantly higher SI than those with 1-2 (d = 1.21), 3-4 (d = 1.48), and 5-8 (d = 0.94) anger attacks in the past month. Furthermore, participants with anger attacks at baseline reported higher SI at each post-baseline visit (both Stages 1 and 2) of EMBARC study (d = 0.39-0.77; all p < .05). Associations between anger attacks and SI were significant even after controlling for irritability, hostility, anxious arousal, depression, suicide propensity, and self-reported pain at baseline and lifetime suicidal tendencies. Similar results were found in participants with aggressive behaviors.

CONCLUSION

Anger attacks in outpatients with MDD may be associated with chronically elevated SI.   Clinical Trials Registration: Establishing Moderators and Biosignatures of Antidepressant Response for Clinical Care for Depression (EMBARC); NCT01407094; https://clinicaltrials.gov/ct2/show/NCT01407094.

A metabolome-wide association study in the general population reveals decreased levels of serum laurylcarnitine in people with depression

Depression constitutes a leading cause of disability worldwide. Despite extensive research on its interaction with psychobiological factors, associated pathways are far from being elucidated. Metabolomics, assessing the final products of complex biochemical reactions, has emerged as a valuable tool for exploring molecular pathways. We conducted a metabolome-wide association analysis to investigate the link between the serum metabolome and depressed mood (DM) in 1411 participants of the KORA (Cooperative Health Research in the Augsburg Region) F4 study (discovery cohort). Serum metabolomics data comprised 353 unique metabolites measured by Metabolon. We identified 72 (5.1%) KORA participants with DM. Linear regression tests were conducted modeling each metabolite value by DM status, adjusted for age, sex, body-mass index, antihypertensive, cardiovascular, antidiabetic, and thyroid gland hormone drugs, corticoids and antidepressants. Sensitivity analyses were performed in subcohorts stratified for sex, suicidal ideation, and use of antidepressants. We replicated our results in an independent sample of 968 participants of the SHIP-Trend (Study of Health in Pomerania) study including 52 (5.4%) individuals with DM (replication cohort). We found significantly lower laurylcarnitine levels in KORA F4 participants with DM after multiple testing correction according to Benjamini/Hochberg. This finding was replicated in the independent SHIP-Trend study. Laurylcarnitine remained significantly associated (p value < 0.05) with depression in samples stratified for sex, suicidal ideation, and antidepressant medication. Decreased blood laurylcarnitine levels in depressed individuals may point to impaired fatty acid oxidation and/or mitochondrial function in depressive disorders, possibly representing a novel therapeutic target.

Resting State Functional Connectivity Biomarkers of Treatment Response in Mood Disorders: A Review

There are currently no validated treatment biomarkers in psychiatry. Resting State Functional Connectivity (RSFC) is a popular method for investigating the neural correlates of mood disorders, but the breadth of the field makes it difficult to assess progress toward treatment response biomarkers. In this review, we followed general PRISMA guidelines to evaluate the evidence base for mood disorder treatment biomarkers across diagnoses, brain network models, and treatment modalities. We hypothesized that no treatment biomarker would be validated across these domains or with independent datasets. Results are organized, interpreted, and discussed in the context of four popular analytic techniques: (1) reference region (seed-based) analysis, (2) independent component analysis, (3) graph theory analysis, and (4) other methods. Cortico-limbic connectivity is implicated across studies, but there is no single biomarker that spans analyses or that has been replicated in multiple independent datasets. We discuss RSFC limitations and future directions in biomarker development.

A robust and reproducible connectome fingerprint of ketamine is highly associated with the connectomic signature of antidepressants

Over the past decade, various N-methyl-D-aspartate modulators have failed in clinical trials, underscoring the challenges of developing novel rapid-acting antidepressants based solely on the receptor or regional targets of ketamine. Thus, identifying the effect of ketamine on the brain circuitry and networks is becoming increasingly critical. In this longitudinal functional magnetic resonance imaging study of data from 265 participants, we used a validated predictive model approach that allows the full assessment of brain functional connectivity, without the need for seed selection or connectivity summaries. First, we identified a connectome fingerprint (CFP) in healthy participants (Cohort A, n = 25) during intravenous infusion of a subanesthetic dose of ketamine, compared to normal saline. We then demonstrated the robustness and reproducibility of the discovered ketamine CFP in two separate healthy samples (Cohort B, n = 22; Cohort C, n = 18). Finally, we investigated the ketamine CFP connectivity at 1-week post treatment in major depressive disorder patients randomized to 8 weeks of sertraline or placebo (Cohort D, n = 200). We found a significant, robust, and reproducible ketamine CFP, consistent with reduced connectivity within the primary cortices and within the executive network, but increased connectivity between the executive network and the rest of the brain. Compared to placebo, the ketamine CFP connectivity changes at 1 week predicted response to sertraline at 8 weeks. In each of Cohorts A-C, ketamine significantly increased connectivity in a previously identified antidepressant CFP. Investigating the brain connectivity networks, we successfully identified a robust and reproducible ketamine biomarker that is related to the mechanisms of antidepressants.

Synthesizing independent stagewise trials for optimal dynamic treatment regimes

Dynamic treatment regimes (DTRs) adaptively prescribe treatments based on patients' intermediate responses and evolving health status over multiple treatment stages. Data from sequential multiple assignment randomization trials (SMARTs) are recommended to be used for learning DTRs. However, due to re-randomization of the same patients over multiple treatment stages and a prolonged follow-up period, SMARTs are often difficult to implement and costly to manage, and patient adherence is always a concern in practice. To lessen such practical challenges, we propose an alternative approach to learn optimal DTRs by synthesizing independent trials over different stages. Specifically, at each stage, data from a single randomized trial along with patients' natural medical history and health status in previous stages are used. We use a backward learning method to estimate optimal treatment decisions at a particular stage, where patients' future optimal outcome increments are estimated using data observed from independent trials with future stages' information. Under some conditions, we show that the proposed method yields consistent estimation of the optimal DTRs and we obtain the same learning rates as those from SMARTs. We conduct simulation studies to demonstrate the advantage of the proposed method. Finally, we learn optimal DTRs for treating major depressive disorder (MDD) by stagewise synthesis of two randomized trials. We perform a validation study on independent subjects and show that the synthesized DTRs lead to the greatest MDD symptom reduction compared to alternative methods.

Association between irritability and suicidal ideation in three clinical trials of adults with major depressive disorder

Irritability in pediatric samples is associated with higher rates of subsequent suicide-related outcomes. No study, to date, has evaluated the longitudinal association between irritability and suicidal ideation (SI) in adults with major depressive disorder (MDD). This report evaluated whether irritability is associated with SI at the same visit (i.e., concurrently) and whether early changes in irritability with antidepressant treatment predict subsequent levels of SI. Participants of Combining Medications to Enhance Depression Outcomes (CO-MED, n = 665), Establishing Moderators and Biosignatures of Antidepressant Response in Clinical Care (EMBARC, n = 296), and Suicide Assessment Methodology Study (SAMS, n = 266) were included. Repeated-measures mixed model analyses evaluated concurrent association throughout the trial between irritability (five-item irritability domain of Concise Associated Symptom Tracking scale) and SI (three-item suicidal thoughts factor of Concise Health Risk Tracking scale) after controlling for overall depression (excluding suicidality-related item), and predicted subsequent levels of SI (repeated observations from week-2-to-week-8) based on early (baseline-to-week-2) changes in irritability after controlling for early changes in overall depression. Higher irritability was associated with higher SI concurrently; estimates (standard error) were 0.18 (0.02, p < 0.0001), 0.64 (0.02, p < 0.0001), and 0.26 (0.04, p < 0.0001) in CO-MED, EMBARC, and SAMS respectively. Greater baseline-to-week-2 reductions in irritability predicted lower levels of subsequent SI; estimates (standard errors) were -0.08 (0.03, p = 0.023), -0.50 (0.05, p < 0.0001), and -0.12 (0.05, p = 0.024) in CO-MED, EMBARC, and SAMS, respectively. Controlling for anxiety or insomnia produced similar results. In conclusion, irritability and SI were consistently linked in adults with MDD. These findings support careful assessment of irritability in suicide risk assessment.

Peripheral Markers of Depression

Major Depressive Disorder (MDD) is a leading cause of disability worldwide, creating a high medical and socioeconomic burden. There is a growing interest in the biological underpinnings of depression, which are reflected by altered levels of biological markers. Among others, enhanced inflammation has been reported in MDD, as reflected by increased concentrations of inflammatory markers-C-reactive protein, interleukin-6, tumor necrosis factor-α and soluble interleukin-2 receptor. Oxidative and nitrosative stress also plays a role in the pathophysiology of MDD. Notably, increased levels of lipid peroxidation markers are characteristic of MDD. Dysregulation of the stress axis, along with increased cortisol levels, have also been reported in MDD. Alterations in growth factors, with a significant decrease in brain-derived neurotrophic factor and an increase in fibroblast growth factor-2 and insulin-like growth factor-1 concentrations have also been found in MDD. Finally, kynurenine metabolites, increased glutamate and decreased total cholesterol also hold promise as reliable biomarkers for MDD. Research in the field of MDD biomarkers is hindered by insufficient understanding of MDD etiopathogenesis, substantial heterogeneity of the disorder, common co-morbidities and low specificity of biomarkers. The construction of biomarker panels and their evaluation with use of new technologies may have the potential to overcome the above mentioned obstacles.

Learning Individualized Treatment Rules for Multiple-Domain Latent Outcomes

For many mental disorders, latent mental status from multiple-domain psychological or clinical symptoms may perform as a better characterization of the underlying disorder status than a simple summary score of the symptoms, and they may also serve as more reliable and representative features to differentiate treatment responses. Therefore, in order to address the complexity and heterogeneity of treatment responses for mental disorders, we provide a new paradigm for learning optimal individualized treatment rules (ITRs) by modeling patients' latent mental status. We first learn the multi-domain latent states at baseline from the observed symptoms under a restricted Boltzmann machine (RBM) model, through which patients' heterogeneous symptoms are represented using an economical number of latent variables and yet remains flexible. We then optimize a value function defined by the latent states after treatment by exploiting a transformation of the observed symptoms based on the RBM without modeling the relationship between the latent mental states before and after treatment. The optimal treatment rules are derived using a weighted large margin classifier. We derive the convergence rate of the proposed estimator under the latent models. Simulation studies are conducted to test the performance of the proposed method. Finally, we apply the developed method to real world studies and we demonstrate the utility and advantage of our method in tailoring treatments for patients with major depression, and identify patient subgroups informative for treatment recommendations.

Pretreatment Reward Sensitivity and Frontostriatal Resting-State Functional Connectivity Are Associated With Response to Bupropion After Sertraline Nonresponse

BACKGROUND

Standard guidelines recommend selective serotonin reuptake inhibitors as first-line antidepressants for adults with major depressive disorder, but success is limited and patients who fail to benefit are often switched to non-selective serotonin reuptake inhibitor agents. This study investigated whether brain- and behavior-based markers of reward processing might be associated with response to bupropion after sertraline nonresponse.

METHODS

In a two-stage, double-blinded clinical trial, 296 participants were randomized to receive 8 weeks of sertraline or placebo in stage 1. Individuals who responded continued on another 8-week course of the same intervention in stage 2, while sertraline and placebo nonresponders crossed over to bupropion and sertraline, respectively. Data from 241 participants were analyzed. The stage 2 sample comprised 87 patients with major depressive disorder who switched medication and 38 healthy control subjects. A total of 116 participants with major depressive disorder treated with sertraline in stage 1 served as an independent replication sample. The probabilistic reward task and resting-state functional magnetic resonance imaging were administered at baseline.

RESULTS

Greater pretreatment reward sensitivity and higher resting-state functional connectivity between bilateral nucleus accumbens and rostral anterior cingulate cortex were associated with positive response to bupropion but not sertraline. Null findings for sertraline were replicated in the stage 1 sample.

CONCLUSIONS

Pretreatment reward sensitivity and frontostriatal connectivity may identify patients likely to benefit from bupropion following selective serotonin reuptake inhibitor failures. Results call for a prospective replication based on these biomarkers to advance clinical care.

Improvements in irritability with sertraline versus placebo: Findings from the EMBARC study

BACKGROUND

This report seeks to evaluate improvements in symptoms of irritability with sertraline (a selective serotonin reuptake inhibitor antidepressant) versus placebo.

METHODS

Participants of Establishing Moderators and Biosignatures of Antidepressant Response in Clinical Care (EMBARC) study who were randomized to 8 weeks of treatment with either sertraline or placebo and completed 5-item irritability domain of Concise Associated Symptom Tracking scale (CAST-IRR) at baseline were included (n = 292). Repeated measures mixed model analysis with CAST-IRR as the outcome variable and treatment arm-by-time interaction as the independent variable of interest evaluated whether changes in irritability with treatment differed between sertraline and placebo arms. A separate analysis controlled for levels of overall depression severity (17-item Hamilton Depression Rating Scale). Covariates included age, sex, race, ethnicity, and site.

RESULTS

There was a significant treatment arm-by-time interaction (F = 6.96, df = 6, 1418, p <0.0001) suggesting that changes in irritability with sertraline differed from placebo. The magnitude of reduction in irritability from baseline to week-8 was greater with sertraline than with placebo (Cohen's d effect size = 0.56). After controlling for levels of overall depression severity at each visit, reduction in irritability with time continued to be significant with sertraline but not with placebo.

LIMITATIONS

Secondary analysis, limited generalizability, lack of non-serotonergic antidepressants.

DISCUSSION

There is greater improvement in irritability with sertraline than with placebo. Improvement in irritability with sertraline was independent of its effects on overall depression severity. CLINICALTRIALS.

GOV IDENTIFIER

NCT01407094.

Connectome-based models can predict early symptom improvement in major depressive disorder

BACKGROUND

Major depressive disorder (MDD) is a debilitating mental illness with more than 50% of patients not achieving an adequate response using first-line treatments. Reliable models that predict antidepressant treatment outcome are needed to guide clinical decision making. We aimed to build predictive models of treatment improvement for MDD patients using machine learning approaches based on fMRI resting-state functional connectivity patterns.

METHODS

Resting-state fMRI data were acquired from 192 untreated MDD patients at recruitment, and their severity of depression was assessed by Hamilton Rating Scale for Depression (HAMD) at baseline. Patients were given medication after the initial MR scan and their symptoms were monitored through HAMD for a period of six months. Connectome-based predictive modeling (CPM) algorithms were implemented to predict the improvement in HAMD score at one month from resting-state connectivity at baseline. Additionally, by selectively combining the features from all leave-one-out iterations in the model building stage, we created a consensus model that could be generalized to predict improvement in HAMD score in samples of non-overlapping subjects at different time points.

RESULTS

Using baseline functional connectivity, CPM successfully predicted symptom improvement of depression at one month. In addition, a consensus 'MDD improvement model' could predict symptom improvement for novel individuals at the two-week, one-month, two-month and three-month time points after antidepressant treatment.

CONCLUSIONS

Individual pre-treatment functional brain networks contain meaningful information that can be gleaned to build predictors of treatment outcome. The identified MDD improvement networks could be an appropriate biomarker for predicting individual therapeutic response of antidepressant treatment. Replication and validation using other large datasets will be a key next step before these models can be used in clinical practice.

Positive and negative valence systems in major depression have distinct clinical features, response to antidepressants, and relationships with immunomarkers

BACKGROUND

Heterogeneity in major depressive disorder (MDD) is well recognized but not well understood. Core depressive features are reward and emotional symptoms, which reflect dysfunctions in the positive valence (PV) and negative valence (NV) systems, respectively. This study assessed whether PV and NV systems (based on selected symptoms) were associated with different clinical features, antidepressant response, and levels of immunomarkers in adults with MDD.

METHODS

These analyses used data from combining medications to enhance depression outcomes study (N = 665; n = 166 for immunomarkers). PV and NV symptom scores were extracted from the clinician-rated 30-item Inventory of Depressive Symptomatology. Correlational analyses were conducted.

RESULTS

PV and NV symptom scores were substantially associated with different clinical features. PV symptoms (impaired motivation, impaired energy, and anhedonia) were independently associated with female gender (p < .001), older age (p = .012), and higher cognitive and physical impairment (p < .001) according to the 7-item Cognitive and Physical Functioning Questionnaire. Conversely, NV symptoms (anxiety and interpersonal sensitivity) were independently associated with younger age (p = .013), more anxious comorbidities (p = .001 for generalized anxiety disorder and p = .002 for social phobia) and other commonly associated noncriterion symptoms (p < .001). Overall, PV symptoms were more responsive to antidepressants than NV symptoms (p < .0001; Cohen's d = .455). A PV symptom score was positively correlated with the concentration of three proinflammatory and one anti-inflammatory factor. In contrast, an NV symptom score was negatively associated with only one proinflammatory immunomarker.

CONCLUSIONS

PV and NV system functions appear to be reflected in selected clinical symptoms that differentially relate to other clinical features, treatment outcomes, and immunological function.

Dorsolateral Prefrontal Cortex and Subcallosal Cingulate Connectivity Show Preferential Antidepressant Response in Major Depressive Disorder

BACKGROUND

Major depressive disorder is associated with abnormal connectivity across emotion and reward circuits as well as other established circuits that may negatively impact treatment response. The goal of this study was to perform an exploratory reanalysis of archival data from a clinical trial to identify moderators of treatment outcome of sertraline over placebo.

METHODS

EMBARC (Establishing Moderators and Biosignatures of Antidepressant Response in Clinical Care) study participants completed magnetic resonance imaging before randomization to either sertraline or placebo for 8 weeks (n = 279). Seed-based functional connectivity was computed using 4 bilateral seeds (2 spheres defined bilaterally): amygdala, dorsolateral prefrontal cortex (DLPFC), subcallosal cingulate cortex, and ventral striatum. Functional connectivity maps were generated, principal component analysis was performed, linear mixed effects models were used to determine moderators of treatment outcome, and post hoc analyses were used to determine level of connectivity (low and high, -1 and +1 SD from the mean) that was most sensitive to improved depression severity (baseline to week 8) based on treatment.

RESULTS

Greater mean reduction in the 17-item Hamilton Rating Scale for Depression score by 8 weeks occurred with sertraline relative to placebo when connectivity in the DLPFC was low (3-way interaction test, p = .05). Conditional on low connectivity in the DLPFC and subcallosal cingulate cortex and high connectivity in the ventral striatum and amygdala, there was on average a 4.8-point greater reduction in the 17-item Hamilton Rating Scale for Depression score with sertraline relative to placebo (p = .003).

CONCLUSIONS

The level of functional connectivity seeded in both the DLPFC and the subcallosal cingulate cortex networks may play an important role in identifying a favorable response to sertraline over placebo.

Dissecting the impact of depression on decision-making

BACKGROUND

Cognitive deficits in depressed adults may reflect impaired decision-making. To investigate this possibility, we analyzed data from unmedicated adults with Major Depressive Disorder (MDD) and healthy controls as they performed a probabilistic reward task. The Hierarchical Drift Diffusion Model (HDDM) was used to quantify decision-making mechanisms recruited by the task, to determine if any such mechanism was disrupted by depression.

METHODS

Data came from two samples (Study 1: 258 MDD, 36 controls; Study 2: 23 MDD, 25 controls). On each trial, participants indicated which of two similar stimuli was presented; correct identifications were rewarded. Quantile-probability plots and the HDDM quantified the impact of MDD on response times (RT), speed of evidence accumulation (drift rate), and the width of decision thresholds, among other parameters.

RESULTS

RTs were more positively skewed in depressed v. healthy adults, and the HDDM revealed that drift rates were reduced-and decision thresholds were wider-in the MDD groups. This pattern suggests that depressed adults accumulated the evidence needed to make decisions more slowly than controls did.

CONCLUSIONS

Depressed adults responded slower than controls in both studies, and poorer performance led the MDD group to receive fewer rewards than controls in Study 1. These results did not reflect a sensorimotor deficit but were instead due to sluggish evidence accumulation. Thus, slowed decision-making-not slowed perception or response execution-caused the performance deficit in MDD. If these results generalize to other tasks, they may help explain the broad cognitive deficits seen in depression.

Reward related ventral striatal activity and differential response to sertraline versus placebo in depressed individuals

Medications to treat major depressive disorder (MDD) are not equally effective across patients. Given that neural response to rewards is altered in MDD and given that reward-related circuitry is modulated by dopamine and serotonin, we examined, for the first time, whether reward-related neural activity moderated response to sertraline, an antidepressant medication that targets these neurotransmitters. A total of 222 unmedicated adults with MDD randomized to receive sertraline (n = 110) or placebo (n = 112) in the Establishing Moderators and Biosignatures of Antidepressant Response in Clinical Care (EMBARC) study completed demographic and clinical assessments, and pretreatment functional magnetic resonance imaging while performing a reward task. We tested whether an index of reward system function in the ventral striatum (VS), a key reward circuitry region, moderated differential response to sertraline versus placebo, assessed with the Hamilton Rating Scale for Depression (HSRD) over 8 weeks. We observed a significant moderation effect of the reward index, reflecting the temporal dynamics of VS activity, on week-8 depression levels (Fs ≥ 9.67, ps ≤ 0.002). Specifically, VS responses that were abnormal with respect to predictions from reinforcement learning theory were associated with lower week-8 depression symptoms in the sertraline versus placebo arms. Thus, a more abnormal pattern of pretreatment VS dynamic response to reward expectancy (expected outcome value) and prediction error (difference between expected and actual outcome), likely reflecting serotonergic and dopaminergic deficits, was associated with better response to sertraline than placebo. Pretreatment measures of reward-related VS activity may serve as objective neural markers to advance efforts to personalize interventions by guiding individual-level choice of antidepressant treatment.

Discovery and replication of cerebral blood flow differences in major depressive disorder

Major depressive disorder (MDD) is a serious, heterogeneous disorder accompanied by brain-related changes, many of which are still to be discovered or refined. Arterial spin labeling (ASL) is a neuroimaging technique used to measure cerebral blood flow (CBF; perfusion) to understand brain function and detect differences among groups. CBF differences have been detected in MDD, and may reveal biosignatures of disease-state. The current work aimed to discover and replicate differences in CBF between MDD participants and healthy controls (HC) as part of the EMBARC study. Participants underwent neuroimaging at baseline, prior to starting study medication, to investigate biosignatures in MDD. Relative CBF (rCBF) was calculated and compared between 106 MDD and 36 HC EMBARC participants (whole-brain Discovery); and 58 MDD EMBARC participants and 58 HC from the DLBS study (region-of-interest Replication). Both analyses revealed reduced rCBF in the right parahippocampus, thalamus, fusiform and middle temporal gyri, as well as the left and right insula, for those with MDD relative to HC. Both samples also revealed increased rCBF in MDD relative to HC in both the left and right inferior parietal lobule, including the supramarginal and angular gyri. Cingulate and prefrontal regions did not fully replicate. Lastly, significant associations were detected between rCBF in replicated regions and clinical measures of MDD chronicity. These results (1) provide reliable evidence for ASL in detecting differences in perfusion for multiple brain regions thought to be important in MDD, and (2) highlight the potential role of using perfusion as a biosignature of MDD.

Individual Differences in Response to Antidepressants: A Meta-analysis of Placebo-Controlled Randomized Clinical Trials

IMPORTANCE

Antidepressants are commonly used worldwide to treat major depressive disorder. Symptomatic response to antidepressants can vary depending on differences between individuals; however, this variability may reflect nonspecific or random factors.

OBJECTIVES

To investigate the assumption of systematic variability in symptomatic response to antidepressants and to assess whether this variability is associated with severity of major depressive disorder, antidepressant class, or year of study publication.

DATA SOURCES

Data used were from a recent network meta-analysis of acute treatment with licensed antidepressants in adults with major depressive disorder. The following databases were searched from inception to January 8, 2016: the Cochrane Central Register of Controlled Trials, CINAHL, Embase, LILACS database, MEDLINE, MEDLINE In-Process, and PsycINFO. Additional sources were international trial registries, drug approval agency websites, and key scientific journals.

STUDY SELECTION

Analysis was restricted to double-blind, randomized placebo-controlled trials with available data at the study's end point.

DATA EXTRACTION AND SYNTHESIS

Baseline and end point means, SDs, number of participants in each group, antidepressant class, and publication year were extracted. The data were analyzed between August 14 and November 18, 2019.

MAIN OUTCOMES AND MEASURES

With the use of validated methods, coefficients of variation were derived for antidepressants and placebo, and their ratios were calculated to compare outcome variability between antidepressant and placebo. Ratios were entered into a random-effects model, with the expectation that response to antidepressants would be more variable than response to placebo. Analysis was repeated after stratifying by baseline severity of depression, antidepressant class (selective serotonin reuptake inhibitors: citalopram, escitalopram, fluoxetine, fluvoxamine, paroxetine, sertraline, and vilazodone; serotonin and norepinephrine reuptake inhibitors: desvenlafaxine and venlafaxine; norepinephrine-dopamine reuptake inhibitor: bupropion; noradrenergic agents: amitriptyline and reboxetine; and other antidepressants: agomelatine, mirtazapine, and trazodone), and publication year.

RESULTS

In the 87 eligible randomized placebo-controlled trials (17 540 unique participants), there was significantly more variability in response to antidepressants than to placebo (coefficients of variation ratio, 1.14; 95% CI, 1.11-1.17; P < .001). Baseline severity of depression did not moderate variability in response to antidepressants. Variability in response to selective serotonin reuptake inhibitors was lower than variability in response to noradrenergic agents (coefficients of variation ratio, 0.88; 95% CI, 0.80-0.97; P = .01), as was the variability in response to other antidepressants compared with noradrenergic agents (coefficients of variation ratio, 0.87; 95% CI, 0.79-0.97; P = .001). Variability also tended to be lower in studies that were published more recently, with coefficients of variation changing by a value of 0.005 (95% CI, 0.002-0.008; P = .003) for every year a study is more recent.

CONCLUSIONS AND RELEVANCE

Individual differences may be systematically associated with responses to antidepressants in major depressive disorder beyond placebo effects or statistical factors. This study provides empirical support for identifying moderators and personalizing antidepressant treatment.

Cortical Connectivity Moderators of Antidepressant vs Placebo Treatment Response in Major Depressive Disorder: Secondary Analysis of a Randomized Clinical Trial

IMPORTANCE

Despite the widespread awareness of functional magnetic resonance imaging findings suggesting a role for cortical connectivity networks in treatment selection for major depressive disorder, its clinical utility remains limited. Recent methodological advances have revealed functional magnetic resonance imaging-like connectivity networks using electroencephalography (EEG), a tool more easily implemented in clinical practice.

OBJECTIVE

To determine whether EEG connectivity could reveal neural moderators of antidepressant treatment.

DESIGN, SETTING, AND PARTICIPANTS

In this nonprespecified secondary analysis, data were analyzed from the Establishing Moderators and Biosignatures of Antidepressant Response in Clinic Care study, a placebo-controlled, double-blinded randomized clinical trial. Recruitment began July 29, 2011, and was completed December 15, 2015. A random sample of 221 outpatients with depression aged 18 to 65 years who were not taking medication for depression was recruited and assessed at 4 clinical sites. Analysis was performed on an intent-to-treat basis. Statistical analysis was performed from November 16, 2018, to May 23, 2019.

INTERVENTIONS

Patients received either the selective serotonin reuptake inhibitor sertraline hydrochloride or placebo for 8 weeks.

MAIN OUTCOMES AND MEASURES

Electroencephalographic orthogonalized power envelope connectivity analyses were applied to resting-state EEG data. Intent-to-treat prediction linear mixed models were used to determine which pretreatment connectivity patterns were associated with response to sertraline vs placebo. The primary clinical outcome was the total score on the 17-item Hamilton Rating Scale for Depression, administered at each study visit.

RESULTS

Of the participants recruited, 9 withdrew after first dose owing to reported adverse effects, and 221 participants (150 women; mean [SD] age, 37.8 [12.7] years) underwent EEG recordings and had high-quality pretreatment EEG data. After correction for multiple comparisons, connectome-wide analyses revealed moderation by connections within and between widespread cortical regions-most prominently parietal-for both the antidepressant and placebo groups. Greater alpha-band and lower gamma-band connectivity predicted better placebo outcomes and worse antidepressant outcomes. Lower connectivity levels in these moderating connections were associated with higher levels of anhedonia. Connectivity features that moderate treatment response differentially by treatment group were distinct from connectivity features that change from baseline to 1 week into treatment. The group mean (SD) score on the 17-item Hamilton Rating Scale for Depression was 18.35 (4.58) at baseline and 26.14 (30.37) across all time points.

CONCLUSIONS AND RELEVANCE

These findings establish the utility of EEG-based network functional connectivity analyses for differentiating between responses to an antidepressant vs placebo. A role emerged for parietal cortical regions in predicting placebo outcome. From a treatment perspective, capitalizing on the therapeutic components leading to placebo response differentially from antidepressant response should provide an alternative direction toward establishing a placebo signature in clinical trials, thereby enhancing the signal detection in randomized clinical trials.

TRIAL REGISTRATION

ClinicalTrials.gov identifier: NCT01407094.

An electroencephalographic signature predicts antidepressant response in major depression

Antidepressants are widely prescribed, but their efficacy relative to placebo is modest, in part because the clinical diagnosis of major depression encompasses biologically heterogeneous conditions. Here, we sought to identify a neurobiological signature of response to antidepressant treatment as compared to placebo. We designed a latent-space machine-learning algorithm tailored for resting-state electroencephalography (EEG) and applied it to data from the largest imaging-coupled, placebo-controlled antidepressant study (n = 309). Symptom improvement was robustly predicted in a manner both specific for the antidepressant sertraline (versus placebo) and generalizable across different study sites and EEG equipment. This sertraline-predictive EEG signature generalized to two depression samples, wherein it reflected general antidepressant medication responsivity and related differentially to a repetitive transcranial magnetic stimulation treatment outcome. Furthermore, we found that the sertraline resting-state EEG signature indexed prefrontal neural responsivity, as measured by concurrent transcranial magnetic stimulation and EEG. Our findings advance the neurobiological understanding of antidepressant treatment through an EEG-tailored computational model and provide a clinical avenue for personalized treatment of depression.

Personalized prognostic prediction of treatment outcome for depressed patients in a naturalistic psychiatric hospital setting: A comparison of machine learning approaches

OBJECTIVE

Research on predictors of treatment outcome in depression has largely derived from randomized clinical trials involving strict standardization of treatments, stringent patient exclusion criteria, and careful selection and supervision of study clinicians. The extent to which findings from such studies generalize to naturalistic psychiatric settings is unclear. This study sought to predict depression outcomes for patients seeking treatment within an intensive psychiatric hospital setting and while comparing the performance of a range of machine learning approaches.

METHOD

Depressed patients (N = 484; ages 18-72; 89% White) receiving treatment within a psychiatric partial hospital program delivering pharmacotherapy and cognitive behavioral therapy were split into a training sample and holdout sample. First, within the training sample, 51 pretreatment variables were submitted to 13 machine learning algorithms to predict, via cross-validation, posttreatment Patient Health Questionnaire-9 depression scores. Second, the best performing modeling approach (lowest mean squared error; MSE) from the training sample was selected to predict outcome in the holdout sample.

RESULTS

The best performing model in the training sample was elastic net regularization (ENR; MSE = 20.49, R2 = .28), which had comparable performance in the holdout sample (MSE = 11.26; R2 = .38). There were 14 pretreatment variables that predicted outcome. To demonstrate the translation of an ENR model to personalized prediction of treatment outcome, a patient-specific prognosis calculator is presented.

CONCLUSIONS

Informed by pretreatment patient characteristics, such predictive models could be used to communicate prognosis to clinicians and to guide treatment planning. Identified predictors of poor prognosis may suggest important targets for intervention. (PsycINFO Database Record (c) 2019 APA, all rights reserved).

Anatomically-Informed Data Augmentation for Functional MRI with Applications to Deep Learning

The application of deep learning to build accurate predictive models from functional neuroimaging data is often hindered by limited dataset sizes. Though data augmentation can help mitigate such training obstacles, most data augmentation methods have been developed for natural images as in computer vision tasks such as CIFAR, not for medical images. This work helps to fills in this gap by proposing a method for generating new functional Magnetic Resonance Images (fMRI) with realistic brain morphology. This method is tested on a challenging task of predicting antidepressant treatment response from pre-treatment task-based fMRI and demonstrates a 26% improvement in performance in predicting response using augmented images. This improvement compares favorably to state-of-the-art augmentation methods for natural images. Through an ablative test, augmentation is also shown to substantively improve performance when applied before hyperparameter optimization. These results suggest the optimal order of operations and support the role of data augmentation method for improving predictive performance in tasks using fMRI.

Connectivity of the Cognitive Control Network During Response Inhibition as a Predictive and Response Biomarker in Major Depression: Evidence From a Randomized Clinical Trial

BACKGROUND

In treating major depressive disorder, we lack tests anchored in neurobiology that predict antidepressant efficacy. Cognitive impairments are a particularly disabling feature of major depressive disorder. We tested whether functional connectivity during a response-inhibition task can predict response to antidepressants and whether its changes over time are correlated to symptom changes.

METHODS

We analyzed data from outpatients with major depressive disorder (n = 124) randomized to receive escitalopram, sertraline, or venlafaxine (8 weeks) and healthy control subjects (n = 59; age 18-65 years). Before and after treatment, participants were interviewed and scanned using functional magnetic resonance imaging, and functional connectivity was measured using generalized psychophysiological interaction during response inhibition (Go-NoGo task). We investigated the interaction between treatment type and response (≥50% reduction on self-reported symptoms), coupling differences between responders and nonresponders at baseline, their correlation with symptom improvement, and their changes in time.

RESULTS

During response inhibition, connectivity between the dorsolateral prefrontal cortex/supramarginal gyrus and supramarginal gyrus/middle temporal gyrus was associated with response to sertraline and venlafaxine, but not escitalopram. Sertraline responders had higher functional connectivity between these regions compared with nonresponders, whereas venlafaxine responders had lower functional connectivity. For sertraline, attenuation of connectivity in the precentral and superior temporal gyri correlated with posttreatment symptom improvement. For venlafaxine, enhancement of connectivity between the orbitofrontal cortex and subcortical regions correlated with symptom improvement.

CONCLUSIONS

Connectivity of the cognitive control circuit during response inhibition selectively and differentially predicts antidepressant treatment response and correlates with symptom improvement. These quantitative markers tied to the neurobiology of cognitive features of depression could be used translationally to predict and evaluate treatment response.

Comprehensive phenotyping of depression disease trajectory and risk: Rationale and design of Texas Resilience Against Depression study (T-RAD)

Depression has a chronic and recurrent course often with early onset and is the leading cause of disability worldwide. In contrast to diagnoses for other conditions which rely on precise medical tests, the diagnosis of depression still focuses exclusively on symptom reports. As a result, heterogeneous patient groups are included under broad categories. Furthermore, in the absence of companion diagnostic tests, choosing specific treatments for patients remains imprecise with only one-third of patients entering remission with initial treatment, with others requiring multiple intervention steps to achieve remission. In addition to improving treatment outcomes, disease prevention is essential to reduce overall disease burden. Adolescence is a critical window where complex emotional, social, familial, and biological shifts may predispose to lifelong depression. Thus, personalized medicine, integrating individual variability in genes, brain function, and clinical phenotypes, can offer a comprehensive approach to provide precise diagnosis, novel drug development, optimal treatment assignment, and prevention of illness and its associated burden. Texas Resilience Against Depression study (T-RAD) encompasses two natural history, longitudinal (10 + years), prospective studies (D2K and RAD), each enrolling 2500 participants. The D2K study follows participants (ages 10 years and older) who have a current or past diagnosis of depression or bipolar disorder. The RAD study follows participants aged 10-24 years who are at risk for depression but not yet suffering from the disease. The T-RAD study will help to uncover the socio-demographic, lifestyle, clinical, psychological, and neurobiological factors that contribute to mood disorder onset, recurrence, progression, and differential treatment response.

Neuroimaging as a Tool for Individualized Treatment Choice in Depression: the Past, the Present and the Future

Purpose of Review

This paper aims to review the findings on neuroimaging as a tool for facilitating individualized treatment choice in depression.

Recent Findings

Neuroimaging has allowed the exploration of neural candidates for response biomarkers. In less than two decades, the field has expanded from small single drug studies to large multisite initiatives testing multiple interventions; from simple analytical methods to employing artificial intelligence, with an aim of establishing models based on a variety of data, such as neuroimaging, biological, psychological and clinical measures.

Summary

Neural biomarkers of response may play an important role in treatment response prediction. It seems likely that they will need to be considered together with other types of data in complex models in order to achieve the high accuracy and generalizability of results necessary for clinical use.

Effect of Intrinsic Patterns of Functional Brain Connectivity in Moderating Antidepressant Treatment Response in Major Depression

OBJECTIVE

Major depressive disorder is associated with aberrant resting-state functional connectivity across multiple brain networks supporting emotion processing, executive function, and reward processing. The purpose of this study was to determine whether patterns of resting-state connectivity between brain regions predict differential outcome to antidepressant medication (sertraline) compared with placebo.

METHODS

Participants in the Establishing Moderators and Biosignatures of Antidepressant Response in Clinical Care (EMBARC) study underwent structural and resting-state functional MRI at baseline. Participants were then randomly assigned to receive either sertraline or placebo treatment for 8 weeks (N=279). A region of interest-based approach was utilized to compute functional connectivity between brain regions. Linear mixed-model intent-to-treat analyses were used to identify brain regions that moderated (i.e., differentially predicted) outcomes between the sertraline and placebo arms.

RESULTS

Prediction of response to sertraline involved several within- and between-network connectivity patterns. In general, higher connectivity within the default mode network predicted better outcomes specifically for sertraline, as did greater between-network connectivity of the default mode and executive control networks. In contrast, both placebo and sertraline outcomes were predicted (in opposite directions) by between-network hippocampal connectivity.

CONCLUSIONS

This study identified specific functional network-based moderators of treatment outcome involving brain networks known to be affected by major depression. Specifically, functional connectivity patterns of brain regions between and within networks appear to play an important role in identifying a favorable response for a drug treatment for major depressive disorder.

Early score fluctuation and placebo response in a study of major depressive disorder

Early score fluctuation in double-blind, placebo-controlled studies may affect the reliability of the baseline measurement and adversely affect the eventual study outcome. We examined the effect of early score fluctuation during a 2-week double-blind placebo lead-in period in a phase II, double-blind, placebo-controlled trial of adjunctive s-adenosyl methionine (MSI-195) in MDD subjects who had had an inadequate response to ongoing antidepressant treatment. The overall study failed to meet its specified endpoints. We examined the score trajectories of all placebo-assigned subjects during the double-blind placebo lead-in period and subsequent 6-week treatment period. Placebo-assigned subjects with ≥20% HamD₁₇ or MADRS score fluctuations (improvement or worsening) during the double-blind placebo lead-in period (prior to randomization) had significantly higher rates of placebo response and remission at week 8 compared to subjects with <20% response. A post-hoc analysis of evaluable subjects taken from the ITT population that excluded subjects with ≥20% early score response yielded higher effect sizes for both the HamD₁₇ and MADRS sub-groups and statistical significance for MSI-195 over placebo in the MADRS sub-group (p = 0.012) with an effect size of 0.404. A reliable baseline measure is an asset for signal detection. These post-hoc findings suggest that study designs that anticipate and attempt to manage early response prior to randomization may yield more meaningful outcome data for trials of MDD and possibly other disorders as well.

Increased Cortical Thickness in Patients With Major Depressive Disorder Following Antidepressant Treatment

Background

Considering the slow-acting properties of traditional antidepressants, an important challenge in the field is the identification of early treatment response biomarkers. Reduced cortical thickness has been reported in neuroimaging studies of depression. However, little is known whether antidepressants reverse this abnormality. In this brief report, we investigated early cortical thickness changes following treatment with sertraline compared to placebo.

Methods

Participants (n = 215) with major depressive disorder were randomized to a selective serotonin reuptake inhibitor, sertraline, or to placebo. Structural magnetic resonance imaging scans were acquired at baseline and one week following treatment. Response was defined as at least 50% improvement in Hamilton rating scale for depression score at week 8. In a vertex-wise approach, we examined the effects of treatment, response, and treatment × response.

Results

Following correction for multiple comparisons, we found a significant effect of treatment, with widespread increase in cortical thickness following sertraline compared to placebo. Clusters with increased thickness were found in the left medial prefrontal cortex, right medial and lateral prefrontal cortex, and within the right parieto-temporal lobes. There were no sertraline-induced cortical thinning, and no significant response effects or treatment × response interactions.

Conclusion

Our findings suggest that cortical thickness abnormalities may be responsive to antidepressant treatment. However, a relationship between these early cortical changes and later treatment response was not demonstrated. Future studies would be needed to investigate whether those early effects are maintained at eight weeks and are associated with enhanced response.

An investigation of cortical thickness and antidepressant response in major depressive disorder: A CAN-BIND study report

Major depressive disorder (MDD) is considered a highly heterogeneous clinical and neurobiological mental disorder. We employed a novel layered treatment design to investigate whether cortical thickness features at baseline differentiated treatment responders from non-responders after 8 and 16 weeks of a standardized sequential antidepressant treatment. Secondary analyses examined baseline differences between MDD and controls as a replication analysis and longitudinal changes in thickness after 8 weeks of escitalopram treatment. 181 MDD and 95 healthy comparison (HC) participants were studied. After 8 weeks of escitalopram treatment (10-20 mg/d, flexible dosage), responders (>50% decrease in Montgomery-Åsberg Depression Scale score) were continued on escitalopram; non-responders received adjunctive aripiprazole (2-10 mg/d, flexible dosage). MDD participants were classified into subgroups according to their response profiles at weeks 8 and 16. Baseline group differences in cortical thickness were analyzed with FreeSurfer between HC and MDD groups as well as between response groups. Two-stage longitudinal processing was used to investigate 8-week escitalopram treatment-related changes in cortical thickness. Compared to HC, the MDD group exhibited thinner cortex in the left rostral middle frontal cortex [MNI(X,Y,Z=-29,9,54.5,-7.7); CWP=0.0002]. No baseline differences in cortical thickness were observed between responders and non-responders based on week-8 or week-16 response profile. No changes in cortical thickness was observed after 8 weeks of escitalopram monotherapy. In a two-step 16-week sequential clinical trial we found that baseline cortical thickness does not appear to be associated to clinical response to pharmacotherapy at 8 or 16 weeks.

Predicting Treatment Outcome in Major Depressive Disorder Using Serotonin 4 Receptor PET Brain Imaging, Functional MRI, Cognitive-, EEG-Based, and Peripheral Biomarkers: A NeuroPharm Open Label Clinical Trial Protocol

BACKGROUND

Between 30 and 50% of patients with major depressive disorder (MDD) do not respond sufficiently to antidepressant regimens. The conventional pharmacological treatments predominantly target serotonergic brain signaling but better tools to predict treatment response and identify relevant subgroups of MDD are needed to support individualized and mechanistically targeted treatment strategies. The aim of this study is to investigate antidepressant-free patients with MDD using neuroimaging, electrophysiological, molecular, cognitive, and clinical examinations and evaluate their ability to predict clinical response to SSRI treatment as individual or combined predictors.

METHODS

We will include 100 untreated patients with moderate to severe depression (>17 on the Hamilton Depression Rating Scale 17) in a non-randomized open clinical trial. We will collect data from serotonin 4 receptor positron emission tomography (PET) brain scans, functional magnetic resonance imaging (fMRI), electroencephalogram (EEG), cognitive tests, psychometry, and peripheral biomarkers, before (at baseline), during, and after 12 weeks of standard antidepressant treatment. Patients will be treated with escitalopram, and in case of non-response at week 4 or intolerable side effects, offered to switch to a second line treatment with duloxetine. Our primary outcome (treatment response) is assessed using the Hamilton depression rating subscale 6-item scores at week 8, compared to baseline. In a subset of the patients (n = ~40), we will re-assess the neurobiological response (using PET, fMRI, and EEG) 8 weeks after initiated pharmacological antidepressant treatment, to map neurobiological signatures of treatment responses. Data from matched controls will either be collected or is already available from other cohorts.

DISCUSSION

The extensive investigational program with follow-up in this large cohort of participants provides a unique possibility to (a) uncover potential biomarkers for antidepressant treatment response, (b) apply the findings for future stratification of MDD, (c) advance the understanding of pathophysiological underpinnings of MDD, and (d) uncover how putative biomarkers change in response to 8 weeks of pharmacological antidepressant treatment. Our data can pave the way for a precision medicine approach for optimized treatment of MDD and also provides a resource for future research and data sharing.

CLINICAL TRIAL REGISTRATION

The study was registered at clinicaltrials.gov prior to initiation (NCT02869035; 08.16.2016, URL: https://clinicaltrials.gov/ct2/results?cond=&term=NCT02869035&cntry=&state=&city=&dist=).

A Unique Brain Connectome Fingerprint Predates and Predicts Response to Antidepressants

More than six decades have passed since the discovery of monoaminergic antidepressants. Yet, it remains a mystery why these drugs take weeks to months to achieve therapeutic effects, although their monoaminergic actions are present rapidly after treatment. In an attempt to solve this mystery, rather than studying the acute neurochemical effects of antidepressants, here we propose focusing on the early changes in the brain functional connectome using traditional statistics and machine learning approaches. Capitalizing on three independent datasets (n = 1,261) and recent developments in data and network science, we identified a specific connectome fingerprint that predates and predicts response to monoaminergic antidepressants. The discovered fingerprint appears to generalize to antidepressants with differing mechanism of action. We also established a consensus whole-brain hierarchical connectivity architecture and provided a set of model-based features engineering approaches suitable for identifying connectomic signatures of brain function in health and disease.

•

Machine learning methods were used to fully investigate the brain connectome

•

Network-informed features engineering approaches were proposed

•

A cortical-subcortical hierarchical brain atlas was established

•

A specific connectome signature was found to predict response to antidepressants

Optimising treatment decision rules through generated effect modifiers: a precision medicine tutorial

This tutorial introduces recent developments in precision medicine for estimating treatment decision rules. The objective of these developments is to advance personalised healthcare by identifying an optimal treatment option for each individual patient based on each patient's characteristics. The methods detailed in this tutorial define composite variables from the patient measures that can be viewed as 'biosignatures' for differential treatment response, which we have termed 'generated effect modifiers'. In contrast to most machine learning approaches to precision medicine, these biosignatures are derived from linear and non-linear regression models and thus have the advantage of easy visualisation and ready interpretation. The methods are illustrated using examples from randomised clinical trials.

Frontal theta and posterior alpha in resting EEG: A critical examination of convergent and discriminant validity

Prior research has identified two resting EEG biomarkers with potential for predicting functional outcomes in depression: theta current density in frontal brain regions (especially rostral anterior cingulate cortex) and alpha power over posterior scalp regions. As little is known about the discriminant and convergent validity of these putative biomarkers, a thorough evaluation of these psychometric properties was conducted toward the goal of improving clinical utility of these markers. Resting 71-channel EEG recorded from 35 healthy adults at two sessions (1-week retest) were used to systematically compare different quantification techniques for theta and alpha sources at scalp (surface Laplacian or current source density [CSD]) and brain (distributed inverse; exact low resolution electromagnetic tomography [eLORETA]) level. Signal quality was evaluated with signal-to-noise ratio, participant-level spectra, and frequency PCA covariance decomposition. Convergent and discriminant validity were assessed within a multitrait-multimethod framework. Posterior alpha was reliably identified as two spectral components, each with unique spatial patterns and condition effects (eyes open/closed), high signal quality, and good convergent and discriminant validity. In contrast, frontal theta was characterized by one low-variance component, low signal quality, lack of a distinct spectral peak, and mixed validity. Correlations between candidate biomarkers suggest that posterior alpha components constitute reliable, convergent, and discriminant biometrics in healthy adults. Component-based identification of spectral activity (CSD/eLORETA-fPCA) was superior to fixed, a priori frequency bands. Improved quantification and conceptualization of frontal theta is necessary to determine clinical utility.

Predicting Response to the Antidepressant Bupropion using Pretreatment fMRI

Major depressive disorder is a primary cause of disability in adults with a lifetime prevalence of 6-21% worldwide. While medical treatment may provide symptomatic relief, response to any given antidepressant is unpredictable and patient-specific. The standard of care requires a patient to sequentially test different antidepressants for 3 months each until an optimal treatment has been identified. For 30-40% of patients, no effective treatment is found after more than one year of this trial-and-error process, during which a patient may suffer loss of employment or marriage, undertreated symptoms, and suicidal ideation. This work develops a predictive model that may be used to expedite the treatment selection process by identifying for individual patients whether the patient will respond favorably to bupropion, a widely prescribed antidepressant, using only pretreatment imaging data. This is the first model to do so for individuals for bupropion. Specifically, a deep learning predictor is trained to estimate the 8-week change in Hamilton Rating Scale for Depression (HAMD) score from pretreatment task-based functional magnetic resonance imaging (fMRI) obtained in a randomized controlled antidepressant trial. An unbiased neural architecture search is conducted over 800 distinct model architecture and brain parcellation combinations, and patterns of model hyperparameters yielding the highest prediction accuracy are revealed. The winning model identifies bupropion-treated subjects who will experience remission with the number of subjects needed-to-treat (NNT) to lower morbidity of only 3.2 subjects. It attains a substantially high neuroimaging study effect size explaining 26% of the variance (R² = 0.26) and the model predicts post-treatment change in the 52-point HAMD score with an RMSE of 4.71. These results support the continued development of fMRI and deep learning-based predictors of response for additional depression treatments.

Prospective cohort study of early biosignatures of response to lithium in bipolar-I-disorders: overview of the H2020-funded R-LiNK initiative

Background

Lithium is recommended as a first line treatment for bipolar disorders. However, only 30% of patients show an optimal outcome and variability in lithium response and tolerability is poorly understood. It remains difficult for clinicians to reliably predict which patients will benefit without recourse to a lengthy treatment trial. Greater precision in the early identification of individuals who are likely to respond to lithium is a significant unmet clinical need.

Structure

The H2020-funded Response to Lithium Network (R-LiNK; http://www.r-link.eu.com/) will undertake a prospective cohort study of over 300 individuals with bipolar-I-disorder who have agreed to commence a trial of lithium treatment following a recommendation by their treating clinician. The study aims to examine the early prediction of lithium response, non-response and tolerability by combining systematic clinical syndrome subtyping with examination of multi-modal biomarkers (or biosignatures), including omics, neuroimaging, and actigraphy, etc. Individuals will be followed up for 24 months and an independent panel will assess and classify each participants’ response to lithium according to predefined criteria that consider evidence of relapse, recurrence, remission, changes in illness activity or treatment failure (e.g. stopping lithium; new prescriptions of other mood stabilizers) and exposure to lithium. Novel elements of this study include the recruitment of a large, multinational, clinically representative sample specifically for the purpose of studying candidate biomarkers and biosignatures; the application of lithium-7 magnetic resonance imaging to explore the distribution of lithium in the brain; development of a digital phenotype (using actigraphy and ecological momentary assessment) to monitor daily variability in symptoms; and economic modelling of the cost-effectiveness of introducing biomarker tests for the customisation of lithium treatment into clinical practice. Also, study participants with sub-optimal medication adherence will be offered brief interventions (which can be delivered via a clinician or smartphone app) to enhance treatment engagement and to minimize confounding of lithium non-response with non-adherence.

Conclusions

The paper outlines the rationale, design and methodology of the first study being undertaken by the newly established R-LiNK collaboration and describes how the project may help to refine the clinical response phenotype and could translate into the personalization of lithium treatment.

Electronic supplementary material

The online version of this article (10.1186/s40345-019-0156-x) contains supplementary material, which is available to authorized users.

Brain regulation of emotional conflict predicts antidepressant treatment response for depression

The efficacy of antidepressant treatment for depression is controversial due to the only modest superiority demonstrated over placebo. However, neurobiological heterogeneity within depression may limit overall antidepressant efficacy. We sought to identify a neurobiological phenotype responsive to antidepressant treatment by testing pretreatment brain activation during response to, and regulation of, emotional conflict as a moderator of the clinical benefit of the antidepressant sertraline versus placebo. Using neuroimaging data from a large randomized controlled trial, we found widespread moderation of clinical benefits by brain activity during regulation of emotional conflict, in which greater downregulation of conflict-responsive regions predicted better sertraline outcomes. Treatment-predictive machine learning using brain metrics outperformed a model trained on clinical and demographic variables. Our findings demonstrate that antidepressant response is predicted by brain activity underlying a key self-regulatory emotional capacity. Leveraging brain-based measures in psychiatry will forge a path toward better treatment personalization, refined mechanistic insights and improved outcomes.

Examining raphe-amygdala structural connectivity as a biological predictor of SSRI response

BACKGROUND

Our lab has previously found that structural integrity in tracts from the raphe nucleus (RN) to the amygdala, measured by fractional anisotropy (FA), predicts remission to selective serotonin reuptake inhibitors (SSRIs) in major depressive disorder (MDD). This could potentially serve as a biomarker for remission that can guide clinical decision-making. To enhance repeatability and reproducibility, we replicated our study in a larger, more representative multi-site sample.

METHODS

64 direction DTI was collected in 144 medication-free patients with MDD from the Establishing Moderators and Biosignatures of Antidepressant Response for Clinical Care (EMBARC) study. We performed probabilistic tractography between the RN and bilateral amygdala and hippocampus and calculated weighted FA in these tracts. Patients were treated with either sertraline or placebo, and their change in Hamilton Depression Rating Scale (HDRS) score reported. Pretreatment weighted FA was compared between remitters and nonremitters, and correlation between FA and percent change in HDRS score was assessed. Exploratory moderator and voxel analyses were also performed.

RESULTS

Contrary to our hypotheses, FA was greater in nonremitters than in remitters in RN-left and right amygdala tracts (p = 0.02 and 0.01, respectively). Pretreatment FA between the raphe and left amygdala correlated with greater, not reduced, HDRS (r = 0.18, p = 0.04). This finding was found to be greater in the placebo group. Moderator and voxel analyses yielded no significant findings.

CONCLUSIONS

We found greater FA in nonremitters between the RN and amygdala than in remitters, and a correlation between FA and symptom worsening, particularly with placebo. These findings may help reveal more about the nature of MDD, as well as guide research methods involving placebo response.

Sensitivity of derived clinical biomarkers to rs-fMRI preprocessing software versions

When common software packages (CONN and SPM) are used to process fMRI, results such as functional connectivity measures can substantially differ depending on the versions of the packages used and the tools used to convert image formats such as DICOM to NIFTI. The significance of these differences are illustrated within the context of a realistic research application: finding moderators of antidepressant response from a large psychiatric study of 288 major depressive disorder (MDD) patients. Significant differences in functional connectivity measurements and discrepancies in derived moderators were found between nearly all software configurations. These results should encourage researchers to be vigilant of software versions during fMRI preprocessing, to maintain consistency throughout each project, and to carefully report versions to facilitate reproducibility.

The Canadian Biomarker Integration Network in Depression (CAN-BIND): magnetic resonance imaging protocols

Studies of clinical populations that combine MRI data generated at multiple sites are increasingly common. The Canadian Biomarker Integration Network in Depression (CAN-BIND; www.canbind.ca) is a national depression research program that includes multimodal neuroimaging collected at several sites across Canada. The purpose of the current paper is to provide detailed information on the imaging protocols used in a number of CAN-BIND studies. The CAN-BIND program implemented a series of platform-specific MRI protocols, including a suite of prescribed structural and functional MRI sequences supported by real-time monitoring for adherence and quality control. The imaging data are retained in an established informatics and databasing platform. Approximately 1300 participants are being recruited, including almost 1000 with depression. These include participants treated with antidepressant medications, transcranial magnetic stimulation, cognitive behavioural therapy and cognitive remediation therapy. Our ability to analyze the large number of imaging variables available may be limited by the sample size of the substudies. The CAN-BIND program includes a multimodal imaging database supported by extensive clinical, demographic, neuropsychological and biological data from people with major depression. It is a resource for Canadian investigators who are interested in understanding whether aspects of neuroimaging — alone or in combination with other variables — can predict the outcomes of various treatment modalities.

Sex differences in the association of baseline c-reactive protein (CRP) and acute-phase treatment outcomes in major depressive disorder: Findings from the EMBARC study

Peripheral inflammation is associated with poor response to antidepressant treatments. However, whether sex differentially affects this association remains unknown. Participants of Establishing Moderators and Biosignatures of Antidepressant Response for Clinical Care (EMBARC) with baseline plasma samples were included in this study (n = 220; male n = 75, female n = 145). Depression severity [Hamilton Rating Scale for Depression 17-item (HAMD-17)] was measured at baseline and weeks- 1, 2, 3, 4, 6, and 8. Plasma c-reactive protein (CRP) was measured with commercially-available ELISA kits at baseline, week-1, and week-8. Sex difference in prediction of baseline-to-week-8 HAMD-17 change by baseline CRP was tested with sex-by-baseline-CRP-by-time interaction in mixed model analysis. Additionally, changes in CRP from baseline-to-week-8 CRP and its association with HAMD-17 changes over that period were also evaluated. Covariates included body mass index, site, smoking status, and age. There was a significant sex difference in association of baseline-to-week-8 HAMD-17 reduction with baseline CRP (p = 0.033). Higher baseline CRP was associated with lower baseline-to-week-8 HAMD-17 reduction in females (p < 0.0001) but not in males (p = 0.632). Additionally, CRP was significantly reduced (p = 0.041, effect size = 0.254) from baseline-to-week-8, but there were no sex differences in this reduction (p = 0.249). Baseline-to-week-8 changes in HAMD-17 and CRP were not significantly associated either overall (p = 0.348) or based on sex (p = 0.370). In a large study of depressed outpatients, we replicated previous findings that elevated baseline CRP levels are associated with worse antidepressant treatment outcomes. However, this effect was limited only to females. These findings emphasize the importance of studying sex differences in biological mechanisms linking inflammation and depression.

Machine learning methods for developing precision treatment rules with observational data

Clinical trials have identified a variety of predictor variables for use in precision treatment protocols, ranging from clinical biomarkers and symptom profiles to self-report measures of various sorts. Although such variables are informative collectively, none has proven sufficiently powerful to guide optimal treatment selection individually. This has prompted growing interest in the development of composite precision treatment rules (PTRs) that are constructed by combining information across a range of predictors. But this work has been hampered by the generally small samples in randomized clinical trials and the use of suboptimal analysis methods to analyze the resulting data. In this paper, we propose to address the sample size problem by: working with large observational electronic medical record databases rather than controlled clinical trials to develop preliminary PTRs; validating these preliminary PTRs in subsequent pragmatic trials; and using ensemble machine learning methods rather than individual algorithms to carry out statistical analyses to develop the PTRs. The major challenges in this proposed approach are that treatment are not randomly assigned in observational databases and that these databases often lack measures of key prescriptive predictors and mental disorder treatment outcomes. We proposed a tiered case-cohort design approach that uses innovative methods for measuring and balancing baseline covariates and estimating PTRs to address these challenges.

Genetic endophenotypes for insomnia of major depressive disorder and treatment-induced insomnia

Major depressive disorder (MDD) is primarily hinged on the presence of either low mood and/or anhedonia to previously pleasurable events for a minimum of 2 weeks. Other clinical features that characterize MDD include disturbances in sleep, appetite, concentration and thoughts. The combination of any/both of the primary MDD symptoms as well as any four of the other clinical features has been referred to as MDD. The challenge for replicating gene association findings with phenotypes of MDD as well as its treatment outcome is putatively due to stratification of MDD patients. Likelihood for replication of gene association findings is hypothesized with specificity in symptoms profile (homogenous clusters of symptom/individual symptoms) evaluated. The current review elucidates the genetic factors that have been associated with insomnia symptom of MDD phenotype, insomnia symptom as a constellation of neuro-vegetative cluster of MDD symptom, insomnia symptom of MDD as an individual entity and insomnia feature of treatment outcome. Homozygous CC genotype of 3111T/C, GSK3B-AT/TT genotype of rs33458 and haplotype of TPH1 218A/C were associated with insomnia symptom of MDD. Insomnia symptom of MDD was not resolved in patients with the A/A genotype of HTR2A-rs6311 when treated with SSRI. Homozygous short (SS) genotype-HTTLPR, GG genotype of HTR2A-rs6311 and CC genotype of HTR2A-rs6313 were associated with AD treatment-induced insomnia, while val/met genotype of BDNF-rs6265 and the TT genotype of GSK-3beta-rs5443 reduced it. Dearth of association studies may remain the bane for the identification of robust genetic endophenotypes in line with findings for genotypes of HTR2A-rs6311.

Pretreatment Rostral Anterior Cingulate Cortex Connectivity With Salience Network Predicts Depression Recovery: Findings From the EMBARC Randomized Clinical Trial

BACKGROUND

Baseline rostral anterior cingulate cortex (rACC) activity is a well-replicated nonspecific predictor of depression improvement. The rACC is a key hub of the default mode network, which prior studies indicate is hyperactive in major depressive disorder. Because default mode network downregulation is reliant on input from the salience network and frontoparietal network, an important question is whether rACC connectivity with these systems contributes to depression improvement.

METHODS

Our study evaluated this hypothesis in outpatients (N = 238; 151 female) enrolled in the Establishing Moderators and Biosignatures of Antidepressant Response for Clinical Care (EMBARC) 8-week randomized clinical trial of sertraline versus placebo for major depressive disorder. Depression severity was measured using the Hamilton Rating Scale for Depression, and electroencephalography was recorded at baseline and week 1. Exact low-resolution electromagnetic tomography was used to compute activity from the rACC, and key regions within the default mode network (posterior cingulate cortex), frontoparietal network (left dorsolateral prefrontal cortex), and salience network (right anterior insula [rAI]). Connectivity in the theta band (4.5-7 Hz) and beta band (12.5-21 Hz) was computed using lagged phase synchronization.

RESULTS

Stronger baseline theta-band rACC-rAI (salience network hub) connectivity predicted greater depression improvement across 8 weeks of treatment for both treatment arms (B = -0.57, 95% confidence interval = -1.07, -0.08, p = .03). Early increases in theta-band rACC-rAI connectivity predicted greater likelihood of achieving remission at week 8 (odds ratio = 2.90, p = .03).

CONCLUSIONS

Among patients undergoing treatment, theta-band rACC-rAI connectivity is a prognostic, albeit treatment-nonspecific, indicator of depression improvement, and early connectivity changes may predict clinically meaningful outcomes.

Personalized prediction of antidepressant v. placebo response: evidence from the EMBARC study

BACKGROUND

Major depressive disorder (MDD) is a highly heterogeneous condition in terms of symptom presentation and, likely, underlying pathophysiology. Accordingly, it is possible that only certain individuals with MDD are well-suited to antidepressants. A potentially fruitful approach to parsing this heterogeneity is to focus on promising endophenotypes of depression, such as neuroticism, anhedonia, and cognitive control deficits.

METHODS

Within an 8-week multisite trial of sertraline v. placebo for depressed adults (n = 216), we examined whether the combination of machine learning with a Personalized Advantage Index (PAI) can generate individualized treatment recommendations on the basis of endophenotype profiles coupled with clinical and demographic characteristics.

RESULTS

Five pre-treatment variables moderated treatment response. Higher depression severity and neuroticism, older age, less impairment in cognitive control, and being employed were each associated with better outcomes to sertraline than placebo. Across 1000 iterations of a 10-fold cross-validation, the PAI model predicted that 31% of the sample would exhibit a clinically meaningful advantage [post-treatment Hamilton Rating Scale for Depression (HRSD) difference ⩾3] with sertraline relative to placebo. Although there were no overall outcome differences between treatment groups (d = 0.15), those identified as optimally suited to sertraline at pre-treatment had better week 8 HRSD scores if randomized to sertraline (10.7) than placebo (14.7) (d = 0.58).

CONCLUSIONS

A subset of MDD patients optimally suited to sertraline can be identified on the basis of pre-treatment characteristics. This model must be tested prospectively before it can be used to inform treatment selection. However, findings demonstrate the potential to improve individual outcomes through algorithm-guided treatment recommendations.

Machine learning and big data in psychiatry: toward clinical applications

Psychiatry is a medical field concerned with the treatment of mental illness. Psychiatric disorders broadly relate to higher functions of the brain, and as such are richly intertwined with social, cultural, and experiential factors. This makes them exquisitely complex phenomena that depend on and interact with a large number of variables. Computational psychiatry provides two ways of approaching this complexity. Theory-driven computational approaches employ mechanistic models to make explicit hypotheses at multiple levels of analysis. Data-driven machine-learning approaches can make predictions from high-dimensional data and are generally agnostic as to the underlying mechanisms. Here, we review recent advances in the use of big data and machine-learning approaches toward the aim of alleviating the suffering that arises from psychiatric disorders.

Cerebral Blood Perfusion Predicts Response to Sertraline versus Placebo for Major Depressive Disorder in the EMBARC Trial

BACKGROUND

Major Depressive Disorder (MDD) has been associated with brain-related changes. However, biomarkers have yet to be defined that could "accurately" identify antidepressant-responsive patterns and reduce the trial-and-error process in treatment selection. Cerebral blood perfusion, as measured by Arterial Spin Labelling (ASL), has been used to understand resting-state brain function, detect abnormalities in MDD, and could serve as a marker for treatment selection. As part of a larger trial to identify predictors of treatment outcome, the current investigation aimed to identify perfusion predictors of treatment response in MDD.

METHODS

For this secondary analysis, participants include 231 individuals with MDD from the EMBARC study, a randomised, placebo-controlled trial investigating clinical, behavioural, and biological predictors of antidepressant response. Participants received sertraline (n = 114) or placebo (n = 117) and response was monitored for 8 weeks. Pre-treatment neuroimaging was completed, including ASL. A whole-brain, voxel-wise linear mixed-effects model was conducted to identify brain regions in which perfusion levels differentially predict (moderate) treatment response. Clinical effectiveness of perfusion moderators was investigated by composite moderator analysis and remission rates. Composite moderator analysis combined the effect of individual perfusion moderators and identified which contribute to sertraline or placebo as the "preferred" treatment. Remission rates were calculated for participants "accurately" treated based on the composite moderator (lucky) versus "inaccurately" treated (unlucky).

FINDINGS

Perfusion levels in multiple brain regions differentially predicted improvement with sertraline over placebo. Of these regions, perfusion in the putamen and anterior insula, inferior temporal gyrus, fusiform, parahippocampus, inferior parietal lobule, and orbital frontal gyrus contributed to sertraline response. Remission rates increased from 37% for all those who received sertraline to 53% for those who were lucky to have received it and sertraline was their perfusion-preferred treatment.

INTERPRETATION

This large study showed that perfusion patterns in brain regions involved with reward, salience, affective, and default mode processing moderate treatment response favouring sertraline over placebo. Accurately matching patients with defined perfusion patterns could significantly increase remission rates.

FUNDING

National Institute of Mental Health, the Hersh Foundation, and the Center for Depression Research and Clinical Care, Peter O'Donnell Brain Institute at UT Southwestern Medical Center.Trial Registration.Establishing Moderators and Biosignatures of Antidepressant Response for Clinical Care for Depression (EMARC) Registration Number: NCT01407094 (https://clinicaltrials.gov/ct2/show/NCT01407094).

Anxiety and anhedonia in depression: Associations with neuroticism and cognitive control

BACKGROUND

Despite the fact that higher levels of anxiety and anhedonia in Major Depressive Disorder (MDD) are linked to poorer treatment outcomes, mechanisms contributing to these clinical presentations remain unclear. Neuroticism, impaired cognitive control, and blunted reward learning may be critical processes involved in MDD and may help to explain symptoms of anxiety and anhedonia.

METHODS

Using baseline data from patients with early-onset MDD (N = 296) in the Establishing Moderators and Biosignatures of Antidepressant Response in Clinical Care (EMBARC) trial, we conducted a path analysis to model relationships between neuroticism, cognitive control, and reward learning to levels of anxiety and anhedonia.

RESULTS

Neuroticism was positively associated with both anhedonia (standardized coefficient = 0.26, p < .001) and anxiety (standardized coefficient = 0.40, p < .001). Cognitive control was negatively associated with anxiety (standardized coefficient = -0.18, p < .05). Reward learning was not significantly associated with either anxiety or anhedonia.

LIMITATIONS

Extraneous variables not included in the model may have even more influence in explaining symptoms of anxiety and anhedonia. Restricted range in these variables may have attenuated some of the hypothesized relationships. Most important, because this was a cross-sectional analysis in a currently depressed sample, we cannot draw any causal conclusions without experimental and longitudinal data.

CONCLUSIONS

These cross-sectional findings suggest that neuroticism may contribute to anxiety and anhedonia in patients with early onset and either chronic or recurrent MDD, while enhanced cognitive control may protect against anxiety.